Amandeep Vashisht , Stephen J Leisz

Crop phenology of alfalfa and grass pasture fields on the western slope of Colorado, USA, was tracked using an NDVI time series derived from Landsat 7 (ETM+), Landsat 8 (OLI) and Sentinel 2a MSI to obtain high temporal frequency information of crop growth stages. These perennial pastures, which occupy a majority of irrigated land in the region and typically have more than one growing/cutting cycle, are quite diverse in terms of agronomic conditions, and their growth stages do not match average reported, conventional values documented almost two decades ago. Divergence of crop growth from these conventional published values have been recognized for other crop types and regions in the past, and refinement of crop growth calendars to reflect actual regional patterns encouraged. Regional practitioners often turn to the conventional published values for guidance in crop and water management, however we document here that the conventional published values are not applicable for this study region. Besides the climatic differences due to latitudinal and elevation characteristics of the region, anthropogenic interventions like multiple cuttings or grazing in a growing season progressively compound the effect of divergence from the conventional published values in the growing season. Using the data available, we summarize regional crop growth stage lengths, which may be used as a reference in absence of real time remote sensing or other ground information. This study serves as an example of the application of a combination of freely available Landsat and Sentinel satellite data to monitor crop calendars in a region of interest, in order to enhance crop management. The results have an associated uncertainty due to limited temporal coverage of cloud-free satellite images, and will benefit from further improvements such as using increased sites, more years of data, and additional satellites (e.g., addition of Sentinel 2b) for increased temporal frequency.